OSN andSource Camera Identification
Ellison [7] has defined social network as a web-based service that creates connections between users. Social network users are required to create a user profile using their own personal details. According to Alexa [8], excessive usage of social network has been recorded and is continuously growing. This has resulted in the increase of digital crime [9, 10] including privacy risk [11], identity theft [12, 13], and emergence of fake profiles [14, 15]. Such crimes have triggered concern among academicians to discuss security and privacy issues, especially on OSN users.
OSNs hold an abundance of valuable user information (for example, digital photos, videos, user friend list with ID, and table of login and IP data [16]). Users tend to share their photos as a channel to express their feelings and as a starting point to start conversations. Over 300 million OSN users (for instance, Facebook) upload photos and videos that are widely accessible all over the internet. This allows criminals to easily access user information, current status, and posted photos and videos. Typically, criminals tend to download photos from OSN applications and use the data for illegal activities. A study by Castiglione et al. (2013) found that images downloaded from OSN websites would have undergone changes in image size and resolution. Hence, investigators will face difficulties in determining the origin of the images [17]. Hence, taking into consideration the underlying issues, an improved approach of digital image source identification is necessary to be in-line with the fast-growing OSN applications. Digital image forensic practitioners are working towards developing the best approach to solve this issue.
Each digital image contains a random feature and pattern noise that depends on the sensor noise used to shoot the image. This paper aims to identify the camera used to generate a digital picture by examining the image’s texture feature. Based on literature review, there are three main techniques proposed to solve source camera identification problems. These techniques are identification based on sensor pattern noise (SPN), color filter array CFA interpolation, and image statistical feature.
2.1 Image Source Identification Based on SPN